JP2013165780A - Diagnosis support apparatus and method of controlling the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve such a problem that, if the diagnosis name expected by a doctor differs from an estimated diagnosis name, there is a possibility that support information required by the doctor is not presented.SOLUTION: A diagnosis support apparatus acquires (S3000) medical information concerning a diagnosis object as already input information, acquires (S3010) an estimated diagnosis name of a user, acquires (S3050) the influence degree of each piece of medical information, selected based on the already input information, on the expected diagnosis name; selects (S3060) as least a part of the medical information as support information based on the influence degree; and displays (S3070) the support information.

Description

  The present invention relates to a diagnosis support apparatus, and more particularly to a diagnosis support apparatus that provides information for supporting medical diagnosis.

  In the medical field, a doctor displays a medical image obtained by imaging a patient on a monitor, interprets the medical image displayed on the monitor, and observes the state of a lesioned part and changes over time. As an apparatus for generating this type of medical image, for example, an X-ray CT (Computed Tomography) apparatus, an MRI (Magnetic Resonance Imaging) apparatus, an ultrasonic apparatus and the like can be cited. Diagnosis using these medical images (image diagnosis) can be divided into a process of finding abnormal shadows and the like from medical images to be diagnosed and obtaining their characteristics, and a process of differential diagnosis of what the shadows are. it can.

  2. Description of the Related Art Conventionally, for the purpose of assisting a differential diagnosis by a doctor, a medical diagnosis support apparatus has been developed that infers and presents what a shadow is as an input information using abnormal shadow characteristics (interpretation findings) and the like. For example, an apparatus that calculates a probability that a shadow with a chest X-ray CT image is a malignant tumor and a probability that it is a benign tumor and presents the same is considered. Usually, as a correct procedure when such an apparatus is used in an actual clinical site, a differential diagnosis is first performed by a doctor, and then a doctor refers to the inference result output by the medical diagnosis support apparatus as reference information.

  Here, when the medical diagnosis support apparatus presents reference information (inference result) without explanation, there arises a problem that the doctor cannot determine the reliability of the reference information. Therefore, an attempt has been made to provide information for determining whether or not a doctor trusts reference information by presenting information used for derivation of reference information (foundation reasoning). By showing the judgment information for the reference information, the doctor's satisfaction with the presented reference information can be increased. As a result, it can be expected to increase the certainty of diagnosis by the doctor himself.

  For example, Patent Document 1 describes a technique for presenting negative information and positive information with respect to an inference result of an apparatus based on input information (hereinafter referred to as “already input information”). ing. In this technique, information is presented to a diagnosis name (hereinafter referred to as “estimated diagnosis name”) having the largest inference probability (possibility) among the inference results. Further, Patent Literature 1 describes a technique for presenting negative information and positive information for each possible diagnosis name. In this technique, the degree of negation or affirmation of each of the already input information is calculated for each estimated diagnosis name or possible diagnosis name, and negative information and positive information are presented. This makes it possible to present information that has influenced the derivation of the inference result of the device based on the already input information.

  On the other hand, when there is a large amount of uninput information (hereinafter referred to as “non-input information”), there arises a problem that the accuracy of inference by the medical diagnosis support apparatus is lowered. Therefore, an attempt has been made to obtain a more reliable inference result by the apparatus selecting uninput information necessary for inference and prompting the doctor to add the information. By prompting the doctor to confirm uninput information that has a great influence on the inference result, an effect of increasing the certainty of diagnosis by the doctor himself can be expected. Furthermore, an effect of reducing a diagnosis error caused by an oversight of the findings can be expected.

  For example, Patent Document 2 describes a technique for selecting and presenting noticeable non-input information from an inference result based on already input information and an inference result when adding uninput information to already input information. Yes. In this technique, the degree of influence of each non-input information on the inference result is calculated, and the non-input information having a large influence degree is presented. As a method of calculating the degree of influence, a method that focuses on the increase in the inference probability of the estimated diagnosis name (not considering a decrease) and a method that focuses on the sum of the change in the probability of each possible diagnosis name are described. . Thereby, it is possible to present uninput information that has a great influence on the inference result of the apparatus based on the already input information.

Japanese Unexamined Patent Publication No. 2010-200840 Japanese Patent No. 3226400

  Patent Document 1 presents positive information and negative information for each estimated diagnosis name or possible diagnosis name. Further, Patent Literature 2 presents uninput information by paying attention to the change amount of the probability of the estimated diagnosis name and the total change amount of the probability of each diagnosis name. Therefore, there is a case where information that affects a specific diagnosis name other than the estimated diagnosis name is not presented, or information about a large number of diagnosis names is presented in addition to information related to the specific diagnosis name. For this reason, when the diagnosis name assumed by the doctor is different from the estimated diagnosis name, there is a possibility that information required by the doctor may not be presented, or information more than necessary may be presented.

  In view of the above problems, an object of the present invention is to enable efficient presentation of information required by a doctor.

  According to one aspect of the present invention, there is provided a diagnosis support apparatus that provides information for supporting a medical diagnosis, the medical information acquiring means for acquiring medical information related to the target of the medical diagnosis as already input information, and the assumed diagnosis of the user An assumed diagnosis name obtaining means for obtaining a name; an influence degree obtaining means for obtaining a first influence degree indicating an influence degree on the assumed diagnosis name for each of the medical information selected based on the already-input information; Provided is a diagnosis support apparatus comprising: selection means for selecting at least a part of the medical information as support information based on the first degree of influence; and presentation means for presenting the support information. .

  According to the present invention, since information that affects the diagnosis name assumed by the user (doctor) is selected and presented, information required by the user can be efficiently presented.

The figure which shows the apparatus structure of the medical diagnosis assistance apparatus which concerns on 1st Embodiment. The figure which shows the basic composition of the computer which implement | achieves each part of a medical diagnosis assistance apparatus with software. 5 is a flowchart showing an overall processing procedure in the first embodiment. The figure which shows the example of the assistance information in 1st Embodiment. The figure which shows the example of the assistance information in 2nd Embodiment. The figure which shows the example of the assistance information in 3rd Embodiment. The figure showing the example of the correspondence of the state which can be taken with the interpretation findings in 1st Embodiment.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, this invention is not limited to the following embodiment, It shows only the specific example advantageous for implementation of this invention. Moreover, not all combinations of features described in the following embodiments are indispensable for solving the problems of the present invention.

[First Embodiment]
Hereinafter, a diagnosis support apparatus that supports diagnosis based on information associated with a diagnosis name in advance will be described. The medical diagnosis support apparatus according to the first embodiment acquires known medical information related to a case that is a diagnosis target as already input information, and performs diagnosis support related to the case.

  In the following, a case where a medical diagnosis support apparatus is used to acquire a plurality of interpretation findings relating to an abnormal shadow of a lung on a chest X-ray CT image as already input information and present diagnosis support information related to the abnormal shadow is presented. This will be described as an example. Specifically, an inference regarding the type of abnormality (diagnosis name) of the abnormal shadow is performed based on already input information, and diagnosis support information is presented based on the result. Of course, the inference object is not limited to this, and the following diagnosis names and interpretation findings that can be input are merely examples for explaining the processing steps of the medical diagnosis support apparatus.

  FIG. 1 is a diagram illustrating a configuration of a medical diagnosis support apparatus according to the first embodiment. As shown in the figure, the medical diagnosis support apparatus 100 in the present embodiment is connected to a case information input terminal 200.

  The case information input terminal 200 acquires information (such as medical image information and electronic medical record information) related to abnormal lung shadows from a server (not shown) regarding a case to be diagnosed. Then, the medical image is displayed on the monitor in a form that can be interpreted by the user (doctor), and the interpretation findings input by the user are acquired as already input information. In addition, the diagnosis name input by the user as a result of interpretation is acquired as the assumed diagnosis name by the user. In the present embodiment, the user inputs interpretation findings and diagnosis names of medical images displayed on the monitor using a mouse or a keyboard. This process is realized by providing the case information input terminal 200 with a function that can be selected by the GUI using, for example, a template-format interpretation finding input support method. The case information input terminal 200 transmits the already input information related to the abnormal shadow of the lung, the assumed diagnosis name, and the data (such as a representative image) associated therewith to the medical diagnosis support apparatus 100 via the LAN or the like according to the user's request. To do.

  The medical diagnosis support apparatus 100 is configured by components described below.

  The medical information acquisition unit 102 acquires already-input information and abnormal data regarding the abnormal shadow of the lungs input from the case information input terminal 200 to the medical diagnosis support apparatus 100, and includes a support information candidate acquisition unit 106, an inference unit 108, And output to the presentation unit 116.

  The assumed diagnosis name acquisition unit 104 acquires the diagnosis name assumed by the doctor regarding the abnormal shadow of the lungs input from the case information input terminal 200 to the medical diagnosis support apparatus 100 as the assumed diagnosis name, and the influence degree acquisition unit 112 and the presentation To the unit 116.

  The support information candidate acquisition unit 106 acquires (selects) at least one set of medical information determined based on already input information as a support information candidate. The acquired support information candidates are output to the inference unit 108 and the influence level acquisition unit 112.

  The inference unit 108 performs inference based on the already input information acquired by the medical information acquisition unit 102 regarding the abnormal shadow of the lung to be diagnosed, and the probability that the abnormal shadow is the name of each diagnosis (result of inference of already input information) ) Is calculated. Further, inference based on each of the support information candidates acquired by the support information candidate acquisition unit 106 is executed, and the execution result (support information candidate inference result) is acquired. The acquired already input information inference result is output to the comparative diagnosis name acquisition unit 110 and the presentation unit 116. In addition, each support information candidate inference result is output to the influence degree acquisition unit 112.

  The comparative diagnosis name acquisition unit 110 uses the already input information inference result acquired by the inference unit 108 to estimate the diagnosis name (inference diagnosis name) based on the inference of the abnormal shadow. Then, the acquired inference diagnosis name is acquired as a comparative diagnosis name. The acquired comparative diagnosis name is output to the influence level acquisition unit 112 and the presentation unit 116.

  The degree of influence acquisition unit 112 acquires the degree of influence on the assumed diagnosis name (first influence degree) and the degree of influence on the comparative diagnosis name (second influence degree) for each support information candidate. This acquisition is performed based on the support information candidate inference result acquired by the inference unit 108. The acquired influence degree is output to the selection unit 114.

  The selection unit 114 is based on the degree of influence on the assumed diagnosis name (first degree of influence) and the degree of influence on the comparative diagnosis name (second degree of influence) regarding each support information candidate acquired by the influence degree acquisition unit 112. To select the support information to be presented. The selected support information is output to the presentation unit 116.

  The presentation unit 116 generates and displays information to be presented based on the already input information acquired by the medical information acquisition unit 102 and the assumed diagnosis name acquired by the assumed diagnosis name acquisition unit 104. Furthermore, based on the already input information inference result acquired by the inference unit 108, the comparative diagnosis name acquired by the comparative diagnosis name acquisition unit 110, and the support information acquired by the selection unit 114, information to be presented is generated and displayed. To do.

  Note that at least a part of each part of the medical diagnosis support apparatus 100 shown in FIG. 1 may be realized as an independent apparatus. Moreover, you may implement | achieve as software which each implement | achieves a function. In the present embodiment, each unit is assumed to be realized by software.

  FIG. 2 is a diagram illustrating a basic configuration of a computer for realizing the functions of the respective units illustrated in FIG. 1 by executing software. The CPU 1001 mainly controls the operation of each component. The main memory 1002 stores a control program executed by the CPU 1001 or provides a work area when the CPU 1001 executes the program. The magnetic disk 1003 stores an operating system (OS), device drivers for peripheral devices, various application software including programs for performing processing described later, and the like. The display memory 1004 temporarily stores display data generated by the presentation unit 116. The monitor 1005 is, for example, a CRT monitor or a liquid crystal monitor, and displays an image, text, or the like based on data from the display memory 1004. A mouse 1006 and a keyboard 1007 perform pointing input and character input by the user, respectively. The above components are connected to each other via a common bus 1008 so as to communicate with each other.

  Next, the overall processing performed by the medical diagnosis support apparatus 100 will be described using the flowchart of FIG. In this embodiment, it is realized by the CPU 1001 executing a program that realizes the function of each unit stored in the main memory 1002.

  In the following description, each item of interpretation findings is represented by Ij (j = 1 to n), and n types of interpretation findings I1 to In are handled. A state (discrete value) that Ij can take is expressed as Sjk. The range of k varies depending on Ij. In the present embodiment, as an example, it is assumed that an interpretation finding as shown in FIG. 7 can be input, and each interpretation finding can take a state as shown in FIG.

  For example, the “shape” of I1 represents the shape of the abnormal shadow, and takes three states of S11 “spherical”, S12 “branch”, and S13 “irregular”. The “cut” in I2 represents the degree of cut in the abnormal shadow. Further, “roll (blood vessel)” of In represents whether or not a blood vessel is involved in an abnormal shadow.

  In the following description, a set of Ij is expressed as N, and a set of Sjk is expressed as E. However, a plurality of states Sjk of one interpretation finding Ij are not present in one E at the same time. For example, when I1 is S11, S12, and S13 and I2 is S21, S22, S23, and S24, E = {S11, S21} can be taken, but E = {S11, S12} cannot be taken. This is because one interpretation finding takes only one state.

  Here, the set of interpretation findings of the input information is expressed as Nf, and the set of Nf states is expressed as Ef. Further, a subset of Ef is expressed as Efl (l = 1, 2, ...). This Efl corresponds to a candidate for support information. In the following description, the diagnosis name is described using the symbol D. In this embodiment, the primary lung cancer, the lung metastasis of cancer, and other three values are taken as the diagnosis names, and are denoted as D1, D2, and D3, respectively. Here, the inference probability of the diagnosis name Dr (r = 1, 2, 3) when no information is given is expressed as P (Dr) (also referred to as prior probability). Further, the inference probability of the diagnosis name Dr when E is given as input information is expressed as P (Dr | E) (also referred to as posterior probability). Further, in the present embodiment, the most likely diagnosis name (estimated diagnosis name) is used as the comparative diagnosis name, and the assumed diagnosis name is denoted as Ddct and the estimated diagnosis name is denoted as Dinf. Ddct and Dinf take one of D1, D2, and D3.

  Further, the degree of influence of the support information candidate Efl on the assumed diagnosis name Ddct is denoted as IR (Ddct | Efl), and the degree of influence on the estimated diagnosis name Dinf is denoted as IR (Dinf | Efl).

  In step S <b> 3000, the medical diagnosis support apparatus 100 acquires already input information and associated data regarding an abnormal shadow of the lungs input to the medical diagnosis support apparatus 100 as processing by the medical information acquisition unit 102. For example, the interpretation finding information acquired by the medical diagnosis support apparatus 100 in S3000 is I1 “shape”: S11 “spherical”, I3 “radial”: S33 “weak”,..., In “roll (blood vessel)”: Sn3 Suppose that it is "No". In this case, the set Nf of interpretation findings of the input information is Nf = {I1, I3, ..., In}, and the set Ef of Nf states is Ef = {S11, S33, ..., Sn3} It becomes.

  In step S <b> 3010, the medical information acquisition unit 102 acquires the diagnosis name assumed by the doctor regarding the abnormal shadow of the lungs input to the medical diagnosis support apparatus 100 as the assumed diagnosis name. For example, when the diagnosis name acquired by the medical diagnosis support apparatus 100 in S3010 is D3 “Other”, the assumed diagnosis name Ddct is Ddct = D3. In addition, when the input from the doctor acquired by the medical diagnosis support apparatus 100 is information indicating the certainty of the doctor with respect to each diagnosis name such as “possible”, “suspect”, “strongly doubt” The diagnosis name with the highest certainty is selected as the assumed diagnosis name.

  In step S3020, the support information candidate acquisition unit 106 acquires at least one set of medical information determined based on already input information as support information candidates. In the present embodiment, a subset of already input information (ie, Efl) selected based on a predetermined rule is set as a support information candidate. For example, when there are 10 elements of the already-input information Ef, 10 support information candidates are acquired when the rule of limiting the elements included in Efl to 1 is used. In addition, when using the rule of limiting the number of elements included to one or two, 55 support information candidates are obtained by adding 45 sets that combine two of the ten to the above ten. To be acquired.

  In S3030, the inference unit 108 infers a probability (an already input information inference result) that each abnormal shadow is a diagnosis name based on the already input information (that is, Ef) regarding the abnormal shadow of the lung acquired in S3000. . Further, the probability (support information candidate inference result) that the abnormal shadow is the diagnosis name is inferred based only on each support information candidate (Efl in this embodiment) acquired in S3020.

  As an inference method at this time, various existing inference methods such as a Bayesian network, a neural network, and a support vector machine can be used. In this embodiment, a Bayesian network is used as an inference method. The Bayesian network is an inference model that uses conditional probabilities, and it is possible to acquire the inference probabilities (that is, posterior probabilities) of each diagnosis name when the already input information is input. In the present embodiment, the respective probabilities of abnormal shadow types D1, D2, and D3 are acquired.

  In S3040, the comparative diagnosis name acquisition unit 110 determines a comparative diagnosis name using the already input information inference result acquired in S3030. In the present embodiment, the probability (estimated diagnosis name) having the highest probability among the probabilities of the abnormal shadow types D1, D2, and D3 is used as the comparative diagnosis name. For example, when D1 = 25.2%, D2 = 42.5%, and D3 = 32.3%, the comparative diagnosis name (estimated diagnosis name) Dinf is Dinf = D2.

  In S3050, the influence level acquisition unit 112 acquires the influence degree (first influence degree) of each support information candidate on the assumed diagnosis name and the influence degree (second influence degree) on the comparison diagnosis name, respectively. . In the present embodiment, each of the influence degree IR (Ddct | Efl) of the support information candidate Efl on the assumed diagnosis name Ddct and the influence degree IR (Dinf | Efl) of the comparison diagnosis name (estimated diagnosis name) Dinf, Calculated by the following formula.

  That is, the difference between the posterior probability when the diagnosis name is inferred using the support information candidate Efl and the prior probability of the diagnosis name is set as the degree of influence. At this time, if the degree of influence is positive, it is considered that the probability has increased because there is a support information candidate Efl. On the other hand, in the case of a negative value, it is considered that the probability has decreased because there is a support information candidate Efl. In the present embodiment, the case where the degree of influence is positive is referred to as positive, and the case where it is negative is referred to as negative. It should be noted that for each diagnosis name, a correspondence table in which the degree of influence of all possible support information candidates is calculated in advance based on Equation (1) is retained, and in the processing of S3050, from the correspondence table The degree of influence may be acquired.

  In S3060, the selection unit 114 is based on the degree of influence (first influence degree) on the assumed diagnosis name of each support information candidate acquired in S3050 and the degree of influence (second influence degree) on the comparative diagnosis name. , Select support information to be presented. For example, only support information candidates that are negative for the assumed diagnosis name are selected as support information. Alternatively, a support information candidate that is positive for one of the assumed diagnosis name and the comparative diagnosis name and negative for the other diagnosis name is selected. Of course, other methods may be used. In the present embodiment, information is selected from the support information candidate Efl having a negative influence degree IR (Ddct | Efl) on the assumed diagnosis name and a positive influence degree IR (Dinf | Efl) on the comparative diagnosis name To do. In addition, among such support information candidates, the support information candidate having the largest IR (Dinf | El) is selected as support information to be presented. That is, the support information candidate having the highest degree of affirming the comparative diagnosis name is selected from the support information candidates that are positive for the comparative diagnosis name and negative for the assumed diagnosis name.

In S3070, the presentation unit 116 displays the following information on the monitor 1005.
Information related to the abnormal shadow of the lung obtained in S3000 (already input information and representative image).
-The assumed diagnosis name obtained in S3010.
The already input information inference result acquired in S3030.
-The comparative diagnosis name obtained in S3040.
-Support information obtained in S3060.

  FIG. 4 shows an example of the presentation information displayed on the monitor 1005 in the present embodiment. The presentation information 400 includes the representative image 4000 of the abnormal shadow of the lung, the already input information 4010 of the abnormal shadow of the lung acquired in S3000, and the assumed diagnosis name 4020 acquired in S3010. The presentation information 400 includes the already input information inference result 4030 acquired in S3030 and the comparative diagnosis name (estimated diagnosis name) 4040 acquired in S3040. In the example shown in the figure, as the already-input information inference result 4030, the inference probability 4031 of primary lung cancer, the inference probability 4032 of cancer lung metastasis in the already-input information inference result, and other inference probabilities 4033 are displayed in a pie chart. Is done. The presentation information 400 includes support information 4050 acquired in 4060.

  When the assumed diagnosis name and the comparative diagnosis name are the same, only positive information for the matched diagnosis name may be selected and presented as support information. Or you may display positive information and negative information, respectively. In that case, display control may be performed so that positive information and negative information are displayed in different display forms. Furthermore, the support information need not be displayed.

  According to the present embodiment, the doctor, together with the inference result (comparison diagnosis name) presented by the diagnosis support apparatus, is negative information with respect to the diagnosis name (assumed diagnosis name) assumed by the doctor himself, and the estimated diagnosis Information positive to the name can be referred to as support information. Therefore, particularly when the assumed diagnosis name and the comparative diagnosis name are different, it is possible to efficiently confirm information on how the diagnosis support apparatus makes such an estimation necessary for the final judgment of the doctor.

(Modification 1 of the first embodiment)
In the above embodiment, in S3040, the diagnosis name (estimated diagnosis name) having the highest inference probability is acquired as the comparative diagnosis name, but the comparative diagnosis name may be acquired by other methods. For example, a diagnosis name having an importance level greater than that of the assumed diagnosis name may be acquired as a comparative diagnosis name. In this case, it is assumed that a predetermined importance level Qr is defined in advance for each diagnosis name Dr. In the processing of S3040, a diagnosis name having a higher importance than the assumed diagnosis name importance Qdoc (that is, satisfying Qdoc <Qr) is acquired as a comparative diagnosis name. For example, in the above example, when Q1 = 0.9, Q2 = 0.5, Q3 = 0.1 and D2 “cancer lung metastasis” is acquired as the assumed diagnosis name, D1 “primary lung cancer” is used as the comparative diagnosis name. To be acquired. When there are a plurality of diagnosis names having a greater importance than the assumed diagnosis name, it is desirable to acquire the diagnosis name having the highest inference probability among them as the comparison diagnosis name. Alternatively, all diagnosis names having a high possibility (inference probability is greater than a predetermined threshold) among the inference results may be acquired as comparative diagnosis names. Or you may acquire the diagnostic name with the highest importance as a comparative diagnostic name.

  As a result, the already input information that is positive for the diagnosis name having higher importance than the assumed diagnosis name of the doctor and negative for the assumed diagnosis name of the doctor is presented. By referring to this information, the doctor can make a diagnosis in consideration of the possibility of a highly important diagnosis name.

(Modification 2 of the first embodiment)
In the above embodiment, only one support information candidate is selected as support information in S3060, but a plurality of support information candidates may be selected as support information. For example, among support conditions, all support information candidates whose influence value is equal to or greater than a predetermined threshold value may be selected as support information. Further, all support information candidates that satisfy the conditions may be selected as support information. According to this, support information having a great influence can be shown to the user without omission.

(Modification 3 of the first embodiment)
In the above embodiment, in S3000, the doctor interprets the medical image displayed by the medical diagnosis support apparatus 100 and acquires the interpretation findings as input information. However, the type of input information and the input information acquisition method are as follows. It is not limited to. For example, past interpretation reports, medical records, and image analysis results for medical images regarding the subject can be used as input information to be processed (existing input information / non-input information). Furthermore, information on medical examination data including other information that can be used for the diagnosis support processing can also be used as target input information.

[Second Embodiment]
The medical diagnosis support apparatus according to the present embodiment selects and presents medical information (non-input information) that can be input other than already input information as support information candidates.

  The configuration of the medical diagnosis support apparatus according to this embodiment is the same as that in FIG. 1 in the first embodiment. However, the point that the support information candidate acquired by the support information candidate acquisition unit 106 is selected from the non-input information other than the already input information is different from the first embodiment. Moreover, the point which acquires the support information candidate inference result acquired in the inference part 108 based on already input information and a support information candidate differs from 1st Embodiment. Further, the influence degree acquisition process performed by the influence degree acquisition unit 112 is different from the first embodiment in that it is performed based on the already input information inference result acquired by the inference unit 108 and the support information candidate inference result. .

  The basic configuration of a computer that implements the medical diagnosis support apparatus 100 by executing software is the same as that in FIG. 2 in the first embodiment. Further, the flowchart for explaining the overall processing performed by the medical diagnosis support apparatus 100 is the same as that in FIG. However, part of the processing performed by S3020, S3030, S3050, and S3060 is different from that of the first embodiment. Hereinafter, with reference to the flowchart of FIG. 3, only differences from the first embodiment will be described for the entire processing performed by the medical diagnosis support apparatus 100 according to the present embodiment.

  In the following, a set of interpretation findings other than the already-input information Nf is denoted as ¬Nf, and a set whose elements are combinable states in ¬Nf is denoted as Evm (m = 1, 2, ...). This Evm corresponds to a support information candidate in this embodiment. Further, the influence degree of the support information candidate Evm on the assumed diagnosis name Ddct is expressed as IR (Ddct | Evm), and the influence degree on the estimated diagnosis name Dinf is expressed as IR (Dinf | Evm).

  Each process of S3000 and S3010 is the same as each process in the first embodiment.

  In step S3020, the support information candidate acquisition unit 106 acquires at least one set of medical information determined based on already input information as support information candidates. In the present embodiment, unlike the processing in the first embodiment, a set (that is, Evm) generated from uninput information that is medical information other than already input information is acquired as a support information candidate. For example, consider a case where ¬Nf = {I2, I6, I8}, I2 takes S21, S22, S23, S24, I6 takes S61, S62, and I8 takes S81, S82, S83. If the number of elements included in Evm is limited to one, 4 + 2 + 3 = 9 support information candidates are acquired. Further, if the number of elements included is limited to one or two, 9 + 4 × 2 + 4 × 3 + 2 × 3 = 35 support information candidates are acquired as possible combinations.

  In S3030, the inference unit 108 performs the same process as in S3030 of the first embodiment, and acquires the already input information inference result based on the already input information. Further, unlike the processing in the first embodiment, the probability that the abnormal shadow is the diagnosis name for each of the already input information and the support information candidate (Evm in this embodiment) as a set of support information candidate inference results. To infer. This inference can be performed by an inference process similar to that of the first embodiment.

  S3040 is the same as the processing in the first embodiment. That is, an estimated diagnosis name is used as a comparative diagnosis name.

  In S3050, the influence level acquisition unit 112 acquires the influence degree (first influence degree) of each support information candidate on the assumed diagnosis name and the influence degree (second influence degree) on the comparison diagnosis name, respectively. . In the present embodiment, unlike the processing in the first embodiment, the degree of influence IR (Ddct | Evm) on the assumed diagnosis name Ddct of the support information candidate Evm and the degree of influence IR (Dinf | Evm on the estimated diagnosis name Dinf) ) Is calculated by the following equation.

  That is, when the support information candidate Evm is added to the already input information, the degree of influence of Evm is determined to the extent that the inference probability of the target diagnosis name is changed. In other words, when the influence level is positive, the inference probability is increased by adding Evm, and when it is negative, it is considered that the inference probability is decreased. In the present embodiment, as in the first embodiment, the case where the influence degree is positive is referred to as positive, and the case where the influence is negative is referred to as negative.

  In S3060, the selection unit 114 selects support information to be presented based on the degree of influence of each of the support information candidates acquired in S3050 on the assumed diagnosis name and the degree of influence on the comparative diagnosis name. For example, only support information candidates positive for the assumed diagnosis name are selected as support information. Alternatively, a support information candidate that is positive for one of the assumed diagnosis name and the comparative diagnosis name and negative for the other diagnosis name is selected. Of course, other methods may be used. In this embodiment, information is selected from the support information candidate Evm that has a positive impact level IR (Ddct | Evm) on the assumed diagnostic name and a negative impact level IR (Dinf | Evm) on the comparative diagnostic name. To do. In addition, the support information candidate having the largest IR (Ddct | Evm) is selected as support information to be presented. That is, among support information candidates that are positive for the assumed diagnosis name and negative for the comparative diagnosis name, the support information candidate that has the highest degree of affirmation of the assumed diagnosis name is selected.

  S3070 is the same as the processing in the first embodiment.

  FIG. 5 shows an example of the presentation information displayed on the monitor 1005 in the present embodiment. In the example shown in the figure, in addition to the example of FIG. 4, a support information candidate inference result 5060 when support information 4050 is added is included. In the example shown in the figure, the support information candidate inference result 5060 includes a pie chart display of an inference probability 5061 of primary lung cancer, an inference probability 5062 of lung metastasis of cancer, and other inference probabilities 5063 in the already input information inference result. The diagnosis name 5065 when the support information is input is included.

  According to the present embodiment, the doctor has positive information with respect to the diagnosis name (assumed diagnosis name) assumed by the doctor himself, together with the inference result (comparison diagnosis name) presented by the diagnosis support apparatus, and the comparative diagnosis Information that is negative to the name can be referred to as support information. That is, when a value is input, it is possible to refer to information such that the inference result moves away from the comparative diagnosis name and approaches the assumed diagnosis name. Therefore, by checking whether there is support information presented in the image, it is possible to efficiently determine whether the diagnosis name assumed by the doctor is correct.

(Modification 1 of 2nd Embodiment)
In the second embodiment, in S3060, among the support information candidates that have a positive influence on the assumed diagnosis name and a negative influence on the comparative diagnosis name, the influence degree on the assumed diagnosis name is the most. A candidate for support information having a large is selected as support information to be presented. However, the selection condition may be other methods. For example, only the support information that matches the diagnosis name (second comparative diagnosis name) acquired from the support information candidate inference result (second inference result) when the support information is input and the assumed diagnosis name is selected. It may be. That is, by inputting the support information, the one that the estimated diagnosis name changes to the assumed diagnosis name may be selected. In this case, since the estimated diagnosis name changes to the assumed diagnosis name, it is possible to more efficiently determine whether the diagnosis name assumed by the doctor is correct by checking the support information.

(Other modifications of the second embodiment)
The modification described in the first embodiment can also be applied to the second embodiment.

[Third Embodiment]
The medical diagnosis support apparatus according to the present embodiment uses the uninput information as support information candidates and presents the transition of the inference result when the uninput information is input.

  The configuration of the medical diagnosis support apparatus according to this embodiment is the same as that in FIG. 1 in the second embodiment. The basic configuration of a computer that implements the medical diagnosis support apparatus 100 by executing software is the same as that in FIG. 2 in the second embodiment. Further, the flowchart for explaining the overall processing performed by the medical diagnosis support apparatus 100 is the same as that in FIG. However, part of the processing performed by S3020, S3060, and S3070 is different from that of the second embodiment. Hereinafter, with reference to the flowchart of FIG. 3, only the differences from the second embodiment will be described for the entire processing performed by the medical diagnosis support apparatus 100 according to the present embodiment.

  The processes in S3000 and S3010 are the same as the processes in the second embodiment.

  In step S3020, the support information candidate acquisition unit 106 acquires at least one set of medical information determined based on already input as support information candidates. In the present embodiment, unlike the processing in the second embodiment, all sets that can be generated from uninput information (that is, ¬Nf) are acquired as support information candidates. For example, consider a case where ¬Nf = {I2, I6, I8}, I2 takes S21, S22, S23, S24, I6 takes S61, S62, and I8 takes S81, S82, S83. In this case, 9+ {4 × 2 + 4 × 3 + 2 × 3} + 4 × 2 × 3 = 59 support information candidates are acquired as sets that can be generated.

  Each process of S3030 to S3050 is the same as each process in the second embodiment.

  In S3060, the selection unit 114 selects support information to be presented based on the degree of influence of each of the support information candidates acquired in S3050 on the assumed diagnosis name and the degree of influence on the comparative diagnosis name. For example, only support information candidates positive for the comparative diagnosis name are selected as support information. Alternatively, a support information candidate that is positive for one of the assumed diagnosis name and the comparative diagnosis name and negative for the other diagnosis name is selected. Of course, other methods may be used. In this embodiment, information is selected from the support information candidate Evm that has a positive impact level IR (Ddct | Evm) on the assumed diagnostic name and a negative impact level IR (Dinf | Evm) on the comparative diagnostic name. To do. In addition, the support information candidate having the largest IR (Ddct | Evm) is selected as support information to be presented (maximum positive information). Further, information is selected from the support information candidate Evm that has a negative influence degree IR (Ddct | Evm) on the assumed diagnosis name and a positive influence degree IR (Dinf | Evm) on the comparative diagnosis name. In addition, the support information candidate having the smallest IR (Ddct | Evm) is selected as support information to be presented (maximum negative information). That is, among support information candidates that are positive for the assumed diagnosis name and negative for the comparative diagnosis name, the support information candidate that has the highest degree of affirmation of the assumed diagnosis name is selected. Further, among the support information candidates that are negative for the assumed diagnosis name and positive for the comparative diagnosis name, the support information candidate having the highest degree of negation of the assumed diagnosis name is selected.

  In S3070, the presentation unit 116 presents the transition of the inference result when the maximum positive information selected in S3060 and the maximum negative information are input in addition to the processing in the second embodiment.

  FIG. 6 shows an example of the presentation information displayed on the monitor 1005 in the present embodiment. In the example shown in the figure, in addition to the example of FIG. 5, support information candidate inference result transition 6070 in the case of adding maximum positive information and maximum negative information is included. In the example shown in the figure, the support information candidate inference result transition 6070 includes other inference probability markers 6071 in the already input information inference result. Furthermore, other inference probability markers 6072 when the maximum positive information is input and other inference probability markers 6073 when the maximum negative information is input are included. In addition, other inference probabilities 6075 when the maximum affirmative information is input and other inference probabilities 6076 when the maximum negative information is input are included.

  According to the present embodiment, when the doctor considers all possible non-input information along with the inference result (comparison diagnosis name) presented by the diagnosis support, the doctor himself / herself has the diagnosis name (assumed diagnosis name) assumed by the doctor himself / herself. How much the inference probability fluctuates can be referred to as support information. In other words, information on how stable the current inference result can be referred to. Therefore, it is possible to efficiently determine whether the inference result based on the already input information is correct.

(Modification 1 of 3rd Embodiment)
In the third embodiment, in S3020, all sets that can be generated from uninput information (that is, ¬Nf) are acquired as support information candidates. However, not all sets that can be generated need be used as support information candidates. For example, the number of elements included in support information candidates may be limited. In this case, for example, if the number of elements is limited to five, it is possible to take into account fluctuations in the inference result when a maximum of five pieces of information are added to the already input information. In addition, the number of elements included in the support information candidates may be increased stepwise. In this case, first, the element included in the support information candidate is set to 1, and the support information candidate is acquired. Next, the processing of S3030 to S3060 is performed, and maximum positive information and maximum negative information are acquired. And it returns to S3020 and acquires the maximum affirmation support information candidate which added each element of the uninput information which is not selected to the maximum affirmation information. Similarly, a maximum negative information candidate obtained by adding each element of uninput information not selected to the maximum negative information is acquired. Then, the processing from S3030 to S3060 is repeated. This process is repeated until there is uninput information that is not selected in S3020. With this processing, it is possible to consider fluctuations close to fluctuations in the inference results when all combinations are considered while suppressing the calculation cost.

(Modification 2 of the third embodiment)
In the above embodiment, in S3070, the transition of the inference result when the maximum positive information selected in S3060 and the maximum negative information are input is presented. However, the presentation is not necessarily performed. For example, when the difference between the inference result when the maximum affirmative information is input and the inference result when the maximum negative information is input is smaller than the threshold value, a display indicating “stable” may be performed. Further, in such a case, it is not necessary to select the maximum positive information and the maximum negative information in S3060.

(Other modifications of the third embodiment)
The modifications described in the first and second embodiments can also be applied to the third embodiment.

(Other embodiments)
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed. In this case, the program and the storage medium storing the program constitute the present invention.

Claims (27)

A diagnosis support apparatus that supports diagnosis based on information associated with a diagnosis name in advance,
An acquisition means for acquiring a diagnosis name by a user;
Providing means for providing negative information for the diagnosis name by the user based on the information;
A diagnostic support apparatus comprising: The acquisition means further acquires information by a user,
The providing means includes:
Including inference means for obtaining a diagnosis name as an inference result based on information by the user;
The diagnosis support apparatus according to claim 1, wherein negative information is provided for the inference result that is the same as a diagnosis name by the user. The inference means obtains a plurality of inference results,
The providing means provides information that is negative with respect to an inference result identical to the diagnosis name and positive with respect to an inference result different from the diagnosis name from among information by the user. The diagnosis support apparatus according to claim 2. The inference means obtains a plurality of inference results,
The providing means is information that is negative for the same inference result as the diagnosis name and positive information for the inference result different from the diagnosis name, and that is not used for the inference among the inputable information The diagnosis support apparatus according to claim 2, wherein the diagnosis support apparatus is provided.   The diagnosis support apparatus according to claim 3 or 4, wherein the inference result different from the diagnosis name is the same inference result as a diagnosis name having higher importance than the inference result same as the diagnosis name. Acquisition means for acquiring at least a plurality of pieces of information that can be input and a diagnosis name by the user;
Inference means for obtaining a plurality of inference results based on the plurality of information;
For the inference result identical to the diagnosis name and the inference result different from the diagnosis name in the inference means, one of positive and negative information is provided from the inputable information. Providing means;
A diagnostic support apparatus comprising:   The diagnosis support apparatus according to claim 6, wherein the providing unit provides the information that can be input from information used for the inference.   The diagnosis support apparatus according to claim 6, wherein the providing unit provides information that is not used for the inference among the inputable information.   9. The diagnosis support apparatus according to claim 7, wherein the inference result different from the diagnosis name is the same inference result as a diagnosis name having a higher importance than the inference result same as the diagnosis name by the user. .   10. The display control unit according to claim 6, further comprising a display control unit configured to display information for affirming and denying the diagnosis name provided by the providing unit on the display unit in different display forms. The diagnosis support device according to item. A diagnosis support apparatus that provides information for supporting medical diagnosis,
Medical information acquisition means for acquiring medical information related to the medical diagnosis target as already input information;
An assumed diagnosis name obtaining means for obtaining an assumed diagnosis name of the user;
Influence level acquisition means for acquiring a first influence level indicating the influence level on the assumed diagnosis name for each piece of medical information selected based on the already input information;
Selection means for selecting at least a part of the medical information as support information based on the first degree of influence;
Presenting means for presenting the support information;
A diagnostic support apparatus comprising: It further has a comparative diagnostic name acquisition means for acquiring a comparative diagnostic name for comparison with the assumed diagnostic name,
The influence degree acquisition unit further acquires a second influence degree indicating an influence degree on the comparative diagnosis name,
The diagnosis support apparatus according to claim 11, wherein the selection unit selects at least a part of the medical information as support information based on the first influence degree and the second influence degree. . An inference means for obtaining an inference result related to the medical diagnosis based on the already-input information;
The diagnosis support apparatus according to claim 12, wherein the comparative diagnosis name is a diagnosis name having a high possibility among the inference results.   The diagnosis support apparatus according to claim 12, wherein the comparative diagnosis name is a diagnosis name having a higher importance than the assumed diagnosis name.   The diagnosis support apparatus according to claim 11, wherein the support information is at least a part of the already input information.   The diagnosis support apparatus according to claim 11, wherein the support information is at least a part of uninput information that is medical information other than the already input information.   13. The support information according to claim 12, wherein the support information is positive information for one of the assumed diagnosis name and the comparative diagnosis name, and is negative information for the other diagnosis name. Diagnosis support device.   15. The diagnosis support apparatus according to claim 14, wherein the support information is negative information for the assumed diagnosis name and positive information for the comparative diagnosis name.   The diagnosis support information according to claim 16, wherein the support information is information positive for the assumed diagnosis name.   The diagnosis support apparatus according to claim 16, wherein the support information is information positive for the assumed diagnosis name and information negative for the comparative diagnosis name. The inference means obtains a second inference result relating to the medical diagnosis based on each set of the already-input information and the medical information;
The diagnosis support apparatus according to claim 13, wherein the comparative diagnosis name acquisition unit acquires a second comparative diagnosis name from the second inference result.   The diagnosis support apparatus according to claim 21, wherein the support information is information that matches the assumed diagnosis name and the second comparative diagnosis name.   The support information indicates a transition of the inference result when the inference result and the most positive information are input to the assumed diagnosis name and when the most negative information is input to the assumed diagnosis name. The diagnosis support apparatus according to claim 21, wherein the diagnosis support apparatus is correct information. A method for controlling a diagnosis support apparatus that supports diagnosis based on information associated with a diagnosis name in advance,
An acquisition step in which the acquisition means acquires a diagnosis name by the user;
A providing step in which providing means provides negative information for the diagnosis name by the user based on the information;
A method for controlling a diagnosis support apparatus, comprising: A method for controlling a diagnosis support apparatus, comprising:
An acquisition means for acquiring at least a plurality of pieces of information that can be input and a diagnosis name by the user;
An inference step in which an inference means obtains a plurality of inference results based on the plurality of information;
The providing means is configured to input information that is positive for one and negative for the other to the inference result that is the same as the diagnosis name and the inference result different from the diagnosis name in the inference step. A provision step to be provided from within,
A method for controlling a diagnosis support apparatus, comprising: A method for controlling a diagnosis support apparatus that provides information for supporting medical diagnosis,
Medical information acquisition means acquiring medical information related to the medical diagnosis target as already input information;
An assumed diagnosis name obtaining means for obtaining an assumed diagnosis name of the user;
An influence degree acquisition unit acquires a first influence degree indicating an influence degree on the assumed diagnosis name for each piece of medical information selected based on the already-input information;
A selection unit selecting at least a part of the medical information as support information based on the first influence;
Presenting means for presenting the support information;
A method for controlling a diagnosis support apparatus, comprising:   The program for functioning a computer as each means which the diagnosis assistance apparatus of any one of Claims 1 thru | or 23 has.

Images